Pembrolizumab Every 12 Weeks Versus Every 3 Weeks in Treating Patients With Non-small Cell Lung Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pembrolizumab Every 12 Weeks Versus Every 3 Weeks in Treating Patients With Non-small Cell Lung Cancer (NCT04032418), a randomized phase 2 study. The primary outcome is 1-year Progression-free Survival (PFS).

2. Clinical Question

The clinical question concerns treatment frequency rather than a comparison of pembrolizumab with an untreated group or a different medicine. The registry summary frames less frequent administration as a possible way to provide similar disease control with fewer treatments. That stated rationale is not itself evidence that the schedules have similar outcomes. The comparison must be interpreted from the study design and the posted results, including their limited size and the absence of a reported comparative analysis.

The official title describes a multicenter international study. The record lists a randomized phase II design and an enrollment of 9. The registry record does not report the eligibility details or the method used to establish prior clinical benefit in enough detail to reconstruct the exact population. The question addressed by the posted endpoint is therefore narrower than a general question about pembrolizumab: it is a comparison of these two schedules in the population selected for this study.

3. Trial Design

This is a randomized phase II study with a parallel-group structure and no masking. The official title describes it as a multicenter international study. The sponsor is Roswell Park Cancer Institute. The registry lists an enrollment of 9 and identifies lung non-small cell carcinoma as the condition. The record lists the study as active, not recruiting, and gives a primary completion date of 2025-03-14.

Participants were assigned to one of two schedules of the same intervention. Arm I received pembrolizumab 200mg intravenously over 30 minutes every 3 weeks. Arm II received 200mg intravenously over 30 minutes every 12 weeks. In both descriptions, treatment could continue for up to 24 months, unless disease progression or unacceptable toxicity occurred. The registry record does not report additional details about allocation concealment, the randomization sequence, or procedures used to maintain or assess masking; masking is listed as none.

The design makes the dosing interval the central contrast, while the medicine and stated dose are the same in both arms. A parallel design means the listed groups are separate treatment groups, rather than participants switching between the two schedules. The record does not report crossover arrangements. Nor does it report a protocol-defined non-inferiority or equivalence margin, a sample-size justification, or other detail that would establish how the phase II study was intended to support a claim of schedule equivalence.

The primary outcome is labeled 1-year progression-free survival and is recorded as a participant count. The posted arm counts are 4 for the every-3-weeks schedule, with 5 listed in the arm, and 2 for the every-12-weeks schedule, with 3 listed in that arm. One participant did not have an assessment within 12 months and was excluded from the primary endpoint analysis; the record says this participant was not considered progressed. The registry also lists overall survival and incidence of adverse events as secondary outcomes. The posted results include no analyses, so the record does not provide a formal statistical comparison between schedules.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Analysis population for the primary outcome, as defined in the registry: One patient did not have assessment within 12 months, thus this patient is not considered progressed for primary endpoint and is excluded from analysis.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as 1-year Progression-free Survival (PFS), measured over 12 months from the start of first cycle of standard dose-schedule of pembrolizumab treatment.

The primary endpoint is titled 1-year progression-free survival (PFS). Its timeframe is 12 months from the start of the first cycle of the standard dose-schedule of pembrolizumab treatment. The registry describes progression using Response Evaluation Criteria In Solid Tumors Criteria, version 1.0 (RECIST v1.0): a 20% increase in the sum of the longest diameter of target lesions, a measurable increase in a non-target lesion, or the appearance of new lesions.

In general, a progression-free survival endpoint is intended to summarize how long participants remain alive without documented disease progression. It can capture clinically relevant change before an overall survival outcome is available, although the way an endpoint is defined and analyzed depends on the study protocol. Here the posted primary result is a count of participants, not a time-to-event estimate, and the registry record does not report a comparative hazard estimate or confidence interval. The record also does not describe death handling in the primary endpoint definition, so that detail should not be inferred from the title alone.

The record states that one participant lacked an assessment within 12 months. That participant was excluded from the primary endpoint analysis and was not considered progressed. This handling affects which participants are represented in the posted endpoint counts. The record does not report further details about other missing assessments or the assessment schedule. The measure is useful for describing the posted 12-month outcome, but the available counts do not show the timing of progression or quantify uncertainty around a between-arm effect.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical information available for this study is limited. The record identifies a randomized, parallel, unmasked phase II trial and posts counts for the primary endpoint and safety outcomes. It does not report a statistical analysis plan or any analyses. The explanations below distinguish general statistical ideas from details that are actually described in the registry record.

What does randomization contribute?

In general, randomization assigns participants to study groups using a chance-based process. Its purpose is to reduce systematic differences between groups at the start of a trial, so that an outcome comparison can more credibly be related to the assigned strategies. Randomization does not guarantee that groups will be alike in every characteristic, especially when the groups are small. The record says the study is randomized, but does not report the sequence-generation method, allocation concealment, stratification factors, or baseline characteristics. Those details are therefore not reported in the registry record.

The parallel structure means that the study presents separate groups receiving the two dosing schedules. The lack of masking is also explicit. In general, masking can reduce the influence of knowledge about assigned treatment on care, assessment, or reporting. Here, the registry lists masking as none. It does not say whether outcome assessment was independently reviewed or whether any other procedures were used to limit assessment bias. Those details are not reported in the registry record.

What can a count-based outcome show?

The primary result is recorded as a number of participants, rather than as a time-to-event estimate. The posted values are 4 in the every-3-weeks group and 2 in the every-12-weeks group; the listed arm sizes for this endpoint are 5 and 3, respectively. A count gives a direct description of how many participants are represented in the recorded outcome. It does not, by itself, describe the time at which progression occurred, the precision of a difference between arms, or the range of effects compatible with the data.

In general, a confidence interval gives a range of values that are compatible with the data under a specified statistical model and procedure. It helps readers understand precision: a broad interval indicates that the estimate is uncertain over a wide range, whereas a narrower interval indicates greater precision. No confidence interval for the arm comparison is posted. A p-value would address compatibility with a specified null hypothesis under a statistical model, but no comparative test or p-value is reported either. Without those quantities, the numerical counts should not be treated as a conclusive test of a schedule effect.

The registry identifies the primary measure as participants with an event (counts). Its endpoint title is 1-year progression-free survival, while the endpoint description gives criteria for progression. These labels make it especially important to follow the registry's stated outcome definition and not recast the posted counts as a percentage or a time-based estimate. The record does not include a plotted or tabular event definition beyond the posted description, or a detailed analysis rule for the count. Any ambiguity about exactly how a count maps to the PFS label is not resolved in the registry record.

How should the sample-size illustration be understood?

In general, sample-size planning asks how many participants would be needed to detect a target effect with a chosen error rate and power, under specified assumptions. A planning calculation is not a result observed in the trial and is not a measure of the probability that this particular study found a true effect. The sample-size section is an illustration based on the observed effect, not the trial's own design. The stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, a normal approximation, and an effect to detect equal to a stated fraction of the observed effect. For time-to-event tables, the stated quantity is total events; for other tables, it is participants per arm.

Those assumptions describe the illustration only. The registry record does not report that they were the assumptions used to design this trial. Nor does a planning illustration establish that the study had sufficient sample size for its intended question. The record does not report a sample-size justification or a target power for the trial. Planning values should be read as hypothetical requirements under the stated calculation assumptions, not as observed power or proof that the trial can answer an equivalence question.

Why is no formal arm comparison shown?

The statistical display indicates that values are missing or not estimable for a two-arm comparison of this binary outcome. In general, a binary outcome records whether a participant falls into one of two categories, such as an event being recorded or not. A comparative analysis would normally summarize the contrast between arms with an effect estimate and a measure of uncertainty, using a method suited to the endpoint and data. The registry record does not provide such an estimate. No method should be imputed from the fact that the trial was randomized.

Over-dispersion is a general issue in some models for count data: observed variability can exceed what the model assumes, leading to uncertainty estimates that are too small if the model is not adapted. This endpoint, however, is identified as binary, and the record does not report a model-based count analysis or any over-dispersion assessment. It would not be appropriate to claim that over-dispersion was tested or relevant to the sponsor's analysis.

Overall, the statistical record supports a descriptive reading of posted counts, along with a separate planning illustration. It does not report adjusted analyses, a protocol-defined analysis population beyond the stated endpoint exclusion, interim analyses, or multiplicity procedures. Those methods cannot be reconstructed from the outcome values alone.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Arm I (200mg Pembrolizumab 3 Weeks)", "Arm II (200mg Pembrolizumab 12 Weeks)")
n      <- c(5, 3)
events <- c(4, 2)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Arm I (200mg Pembrolizumab 3 Weeks)  5  4  80.0  28.4  99.5  37.6  96.4  35.9  95.7
ARM  Arm II (200mg Pembrolizumab 12 Weeks)  3  2  66.7  9.4  99.2  20.8  93.9  19.4  93.2

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome, 1-year progression-free survival, and identifies overall survival and incidence of adverse events as secondary outcomes. It does not report a multiplicity plan. In general, multiplicity arises when a study tests several outcomes, time points, groups, or hypotheses: the chance of finding at least one apparently positive result can increase when many tests are considered. A prespecified strategy can address this issue, but no such strategy is described in the registry record. The posted results also contain no analyses, so there are no reported secondary-outcome tests to interpret.

The record states that one participant did not have an assessment within 12 months, was not considered progressed, and was excluded from the primary endpoint analysis. That is the only specific missing-assessment handling described for the primary endpoint. The registry record does not report an imputation procedure, a sensitivity analysis, or a detailed set of rules for other missing observations. It also does not report whether all randomized participants were included in any other analysis population.

In general, analysis populations define which participants contribute to a particular estimate. For a randomized study, readers often want to know whether analyses follow assigned groups, include only participants who received treatment, or use another prespecified set. These distinctions can matter when participants are excluded after randomization. Here, the record gives the arm counts used for the primary outcome and the separate adverse-event reporting denominators, but does not explain all population definitions or their relationship. The registry record does not report the full statistical analysis plan.

Because no comparative analysis is posted, there is no reported adjustment for multiple testing, no stated missing-data sensitivity result, and no reported primary-analysis model to evaluate. These absences limit how far the endpoint counts can be interpreted. They do not establish that the trial had no such procedures outside the registry; they mean the procedures are not reported in the registry record.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: atrial fibrillation (20.0%), myocardial infarction (20.0%), sepsis (20.0%), generalized muscle weakness (0.0%), hematoma (0.0%).

The adverse-event display reports serious adverse events for the two arms separately. In the every-3-weeks arm, 2 of 5 participants at risk are listed as affected by a serious adverse event. In the every-12-weeks arm, 0 of 4 are listed as affected. The table also lists other adverse events for 5 of 5 participants in the every-3-weeks group and 0 of 4 in the every-12-weeks group. These are counts as presented; the groups have small denominators, so the figures should not be treated as precise estimates of comparative safety.

The serious-event terms listed include atrial fibrillation, myocardial infarction, and sepsis, each shown as 20.0%, along with generalized muscle weakness and hematoma, each shown as 0.0%. These percentages are reproduced as registry entries and should be read with the event counts and group context. The record does not provide further details about event timing, severity grading, attribution, recurrence, or whether a term occurred in more than one participant.

The safety display lists deaths as 4 in the every-3-weeks group and 2 in the every-12-weeks group. These figures alone do not establish that deaths were caused by pembrolizumab or by the dosing schedule. The registry record does not report cause-specific details alongside these counts. The safety denominators also differ from the primary endpoint denominators, and the record does not explain that difference. These figures describe what is posted; they are not a standalone comparative safety analysis.

13. What the Result Does, and Does Not, Mean

The posted results provide a record of observed counts for a small randomized comparison of two pembrolizumab schedules. For the 1-year progression-free survival outcome, the registry lists 4 participants in the every-3-weeks arm and 2 in the every-12-weeks arm, with endpoint denominators of 5 and 3. It also records a participant who lacked a 12-month assessment and was excluded from the primary endpoint analysis. The study's outcome table and plots present these values descriptively.

These results do not establish that the schedules have the same effect. In general, a conclusion of equivalence requires a prespecified range of differences considered acceptably small, as well as an analysis showing that uncertainty is contained within that range. A non-inferiority conclusion likewise depends on a prespecified margin and an appropriate analysis. The registry record does not report either margin or a comparative analysis. The trial title and summary describe an interest in similar disease control, but that stated aim does not substitute for evidence meeting a specified statistical criterion.

The data also do not establish that the schedules differ. The record contains no reported confidence interval or test for a between-arm effect, and the endpoint groups are small. Without a measure of uncertainty, the counts cannot show how precisely a schedule difference has been estimated. The absence of a reported comparison should not be mistaken for a finding of no difference.

The safety counts likewise require restraint. They list affected participants and deaths, but do not establish the cause of an event or death, and the displayed denominators differ across outcome sections. Overall survival and incidence of adverse events are listed as secondary outcomes, but the record does not post analyses for them. No broader claim about benefit, harm, or the preferred schedule follows from these registry counts alone.

In general, the practical meaning of a study result depends on its population, endpoint definition, design, follow-up, completeness of assessment, and precision. The registry gives some of these details, including the 12-month timeframe and RECIST v1.0 progression criteria, but does not report the full analysis plan or comparative uncertainty. The defensible reading is therefore descriptive: the registry posts arm-level counts and safety figures, while the strength and precision of a schedule comparison remain unquantified in the record.

14. Why This Trial Matters Statistically

This trial is statistically interesting because it compares two frequencies of the same named treatment in a randomized, parallel design. The study question is not whether pembrolizumab is compared with no treatment; it is whether the posted outcomes differ when pembrolizumab is given every 12 weeks rather than every 3 weeks. That distinction makes the choice of estimand and the intended standard for a meaningful difference central to interpretation.

The summary describes the possibility of similar disease control with fewer treatments as a rationale for examining the less frequent schedule. Statistically, however, “similar” needs a precise definition. In general, a non-inferiority or equivalence question requires a justified margin set in advance, an analysis suited to that question, and enough information to assess uncertainty. The registry record does not report a margin, sample-size justification, or comparative analysis. The record therefore allows readers to identify the question, but not to verify that the posted results satisfy a formal similarity criterion.

The small posted endpoint groups also illustrate why counts and uncertainty must be distinguished. Counts are transparent and easy to read, but they provide limited information about how stable an apparent contrast might be. Randomization supports the design, but does not overcome the imprecision that can arise with few participants. The absence of a reported confidence interval or test leaves that uncertainty unquantified in the registry.

The record also highlights the importance of endpoint handling. A participant without a 12-month assessment was excluded and not considered progressed, while the safety section uses different group denominators. These choices affect which observations contribute to each summary. Clear population definitions and sensitivity analyses help readers understand the robustness of results; those details are not reported.

Finally, the sample-size section is a planning illustration based on the observed effect and stated assumptions, not a reconstruction of the trial's original design. Separating planning calculations from observed evidence is important: a planning value cannot make a small posted comparison more precise or establish observed power. The trial is a useful case for reading registry outcomes cautiously and distinguishing a research aim from a statistical conclusion.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary limitation of the posted efficacy information is the very small number of participants represented in the endpoint groups: 5 in the every-3-weeks arm and 3 in the every-12-weeks arm. The registry lists an overall enrollment of 9. One participant lacked an assessment within 12 months and was excluded from the primary endpoint analysis. The record does not report additional detail about the excluded participant or the assessment process.

The endpoint is described as 1-year progression-free survival, but the posted result is a count rather than a time-to-event estimate. The progression definition cites RECIST v1.0 criteria, while the registry does not describe death handling in the primary endpoint definition. It also does not provide comparative effect estimates, confidence intervals, or tests. As a result, the size and precision of any between-arm contrast cannot be assessed from the posted analysis.

The record does not report a sample-size justification, a formal equivalence or non-inferiority margin, an analysis plan, interim analyses, or multiplicity procedures. It also does not report baseline characteristics, detailed eligibility criteria, allocation concealment, or the assessment process for progression. These gaps prevent readers from evaluating some aspects of trial conduct and analysis. Their absence from the registry should not be taken as proof that no procedures existed outside the record.

The safety results have their own limits. The event denominators differ from the primary endpoint denominators, and the record does not explain why. Serious-event terms and death counts are posted without the accompanying detail needed to establish timing, attribution, or cause. The secondary outcomes are listed, but analyses are not posted. Taken together, these limitations make the displayed values descriptive rather than a sufficient basis for a definitive comparison of schedules.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Pembrolizumab Every 12 Weeks Versus Every 3 Weeks in Treating Patients With Non-small Cell Lung Cancer — NCT04032418.
• Registry record last updated 2026-10-22; status last verified 2026-10.
• Results first posted 2026-09-21.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

This registry record describes a randomized, unmasked, parallel phase II study of pembrolizumab every 3 weeks versus every 12 weeks in non-small cell lung cancer. Both arms use 200mg intravenously over 30 minutes, with treatment for up to 24 months in the absence of progression or unacceptable toxicity. The record lists enrollment of 9, identifies Roswell Park Cancer Institute as sponsor, and gives a primary completion date of 2025-03-14. Study status is active, not recruiting.

The primary outcome is 1-year progression-free survival, assessed 12 months from the start of the first treatment cycle. Posted counts are 4 of 5 for the every-3-weeks arm and 2 of 3 for the every-12-weeks arm. One participant without a 12-month assessment was excluded from this analysis and was not considered progressed. Safety figures are also posted, but their denominators differ from the endpoint groups.

The registry does not report a formal comparative analysis, confidence interval, sample-size justification, or equivalence or non-inferiority margin. The available results are therefore best read as small descriptive counts rather than a conclusive statistical comparison.